Morphometry , growth and reproduction of an Atlantic population of the razor clam Ensis macha ( Molina , 1782 ) *

Razor clams are bivalve molluscs with low world landings, but they fetch high prices in international markets. Although not reported in fisheries statistics (FAO, 2000), Ensis macha is one of the most important razor clam species in volume of captures, reaching landings of approximately 6,000 t in 1999 in Chile (Sernapesca, 2000), where beds have begun to show clear signs of overexploitation (Bustos et al., SCI. MAR., 68 (2): 211-217 SCIENTIA MARINA 2004


INTRODUCTION
Razor clams are bivalve molluscs with low world landings, but they fetch high prices in international markets.Although not reported in fisheries statistics (FAO, 2000), Ensis macha is one of the most important razor clam species in volume of captures, reaching landings of approximately 6,000 t in 1999 in Chile (Sernapesca, 2000), where beds have begun to show clear signs of overexploitation (Bustos et al., 1999).This species inhabits subtidal sandy-muddy bottoms along coastal waters of the Atlantic (40-55°S, depths of 0 to 55 m; Lasta et al., 1998) and Pacific (27-55°S, depths of 0 to 13 m; Osorio et al., 1979) coasts of South America (Fig. 1).Previous studies on the growth (Urban, 1996) and reproduction (Avellanal et al., 2002) were conducted on Chilean populations, and no information is available on Atlantic populations.Recent work has demonstrated that the beds from the northern-Patagonian Gulfs (Argentina, 42-43°S) have a potential for exploitation (average density in beds = 51.6 clams/m 2 , average biomass = 1-1.5 kg/m 2 ; mean length of individuals = 124.9-134.2mm) (Ciocco, unpublished).The artisanal shellfishery of northern Argentine Patagonia (San José Gulf and South of San Matías Gulf) is operated by hookah divers.In the last few years, 19-22 vessels 8-10 m in length, equipped with outboard engines up to 110 hp, have operated in the area landing 500-1000 tons of bivalves per year.Traditional target species include the tehuelche scallop (Aequipecten tehuelchus), the ribbed mussel (Aulacomya ater), the common mussel (Mytilus edulis) and the striped clam (Protothaca antiqua).Recently, the razor clam and another clam species, the southern geoduck (Panopea abbreviata), have been added.The interest in the exploitation of the former has increased after samples sent to Chile, Asia and domestic markets were found acceptable for marketing (Lasta et al., 1998).The methods used to capture E. macha so far include grabbing it with long pincers (also employed in Chilean fisheries), seizing it directly with the hand when the sediment is soft and inserting a small metal sphere, welded to the end of an iron rod, into the clam's inhalant siphon and pulling it back when the siphon retracts.Given the lack of regulation, there is a need to establish a management plan for this incipient fishery.In support of this need, the present work reports the first data on the morphometry, growth and reproductive patterns of E. macha from northern Argentine Patagonia.

Sampling
A survey was conducted from September 1999 to August 2000 on E. macha beds from Puerto Lobos (San Matías Gulf, Argentina; 42°00'S-65°04'W) (Fig. 1).Razor clam samples (N ≥ 30 individuals) were collected monthly by SCUBA divers by water jet pumping and were fixed in 5% formaldehyde in seawater immediately after capture.In November and December 1999 and February, May, June and July 2000, sediment samples (one sample of 2000 cm 3 each month without replicates) containing E. macha spat were taken from the seabed, frozen, defrosted a few months later and subsequently examined under a dissecting microscope.

Morphometry
The morphometric variables recorded include length (L), height (H), length along the diagonal axis (D) (Fig. 2), dry weight of both valves (VW), wet weight of the whole specimens (TW) and wet weight of the soft parts (SPW).Since individuals captured by SCUBA divers were fixed in formaldehyde before weighing, measurements of TW and SPW may have been biased.Individuals collected in the sediment samples were not included in the analysis of weight variables because they were subjected to a different method of preservation.Given that logarithmic transformation linearises relationships and stabilises variances (Voight, 1991(Voight, , 1994)), all variables were ln-transformed.Regression lines were fitted in order to establish relationships between ln(L) and other ln-transformed variables and between ln(SPW) and ln(TW).The Student´s t test was performed to establish allometry/isometry in the relationships.For one of the relationships showing a morphometric discontinuity, a piecewise regression analysis was performed running the Simplex algorithm included in Statistica (Statsoft Inc., 1996).

Growth
Growth checks were recognised by topographic discontinuities and change of colour on the valves' surface.Growth was characterised by measuring the distances from each check to the hinge along the line of maximum shell growth (D, Fig. 2).Since this method is rather subjective, growth check measurements were performed independently by two observers on the left valves of 98 and 99 individuals respectively from a sample of 147 clams.All measurements were done with Vernier calipers to the nearest 0.1 mm.Taking into account that the measure of size for razor clams in other studies is L (Henderson and Richardson, 1994;Urban, 1996), growth in D was transformed to L using the regression equation that relates the two variables.The von Bertalanffy (1938) equation was fitted to size-at-growth check data by means of a non-linear regression using the Simplex algorithm (Statsoft Inc., 1996) and considering that "t" is the number of growth checks marked by the animal since the beginning of shell formation, "L t " is the valve's length (L) at growth check "t", L ∞ is the asymptotic length, k is a constant and t 0 is the number of growth checks marked at size 0. Growth parameters were estimated with maximum likelihood methods independently for the data registered by each observer.Comparisons between curves were conducted by contrasting the likelihood ratios with chi-squared tests (Cerrato, 1990).To obtain a validation for the periodicity of growth check deposition, 100 individuals were collected from the bed in November 1999, marked with a razor on the surface of their valves perpendicularly to the anterior margins and returned to the bed.Since the animals were lost at the time of recapture (November 2000), probably owing to fishing by the artisanal fishermen operating in the area, a second sample of 100 individuals were marked and left in the bed for one year until recapture (November 2001).

Reproductive patterns
Three hundred and forty-nine individuals ranging from 50 to 165 mm in L were dissected and tissue samples were taken from the gonads.Sex was determined by observation of gonadal smears under binocular microscope.Once all females had been identified, ovary pieces were stained with Methylen Blue, smeared on glass slides and photographed with a digital camera.The areas of the projections of 100 oocytes on the photo images were measured by pixel counting (Aldus Photo Styler, v2) and the area of the rectangles containing each oocyte were calculated by multiplying the length of the largest axis of the oocyte by that of its largest perpendicular axis (measured with Scion image 4).In view of the high correlation between the two measures of area (R = 0.97), their relationship was estimated by regression.Then, the rectangular areas of at least 50 vitellogenic oocytes per female were calculated and transformed into the projection areas using the regression equation.To obtain an estimation of oocyte size, it was considered that oocytes were perfectly circular, their diameters (d) being related to the projection areas by the equation: From these calculations, monthly size frequency distributions of vitellogenic oocytes were drawn.

Morphometry
The morphometric relationships established in this study, along with some sample descriptive statistics are given in Table 1.The left and right valves of E. macha are symmetrical and have a similar weight.Growth of D relative to L is isometric.In contrast, growth of H relative to L is negatively allometric for individuals with L < 11.23 mm and slightly positively allometric for larger ones.Growth of TW relative to L is isometric.On the other hand, growth of SPW relative to L is negatively allometric and that of VW relative to L is positively allometric.Accordingly, SPW relative to TW is negatively allo-metric, SPW varying between approximately 50 and 70% (mean 60%) of TW.

Growth
Observer 1 noted alternating light concavities and dark convexities on the surfaces of the valves of E. macha along their axes of maximum growth, each concavity being separated from the following convexity by a conspicuous step.For this observer, only these steps represented seasonal growth checks.On the other hand, observer 2 considered that a growth check was any conspicuous topographic discontinuity on the valves' surface.Maximum likelihood growth curves obtained by the two independent observers were significantly different from each other, the one obtained by observer 2 showing greater dispersion (Table 2).Between-observer differences were related to parameters k and t 0 (Table 2).The maximum numbers of growth checks were 14 and 16 for observers 1 and 2 respectively, the modal number of checks being seven for both.Even though validation experiments on the periodicity of deposition of the check marks were conducted by marking 200 individuals, only one of them was recaptured alive after one year.This individual grew from L = 125.7 mm (November 2000) to 130.1 mm (November 2001) and deposited only one check mark (its 7 th one) during this period.The other 199 clams were lost due to unknown reasons, probably including fishing, depredation, self-dispersion and scattering due to stormy weather conditions.214 P.J. BARÓN et al.

Reproductive patterns
Out of a total of 349 clams analysed, 183 were males, 166 were females and none was hermaphrodite.The sex ratios of individuals divided into three size classes were not significantly different to 1:1 (Table 3).Vitellogenic oocytes were found in all female clams and in all months sampled, suggesting that in northern Patagonia there is no resting period.Size of the vitellogenic oocytes was in the range 16-110 µm.Monthly frequency distributions of vitellogenic oocyte diameter are shown in Figure 3.In September of 1999, oocyte size classes between 36-80 µm were evenly represented.In November of the same year, the largest size classes were much less frequent than in September, indicating that partial spawning had occurred.In December, there was a recovery of the largest size classes, two modal components being clearly distinguishable until May 2000; by then the largest size classes had become less abundant.From May to June, there was an abrupt disappearance of the largest modal component, indicating that massive spawning took place in this period.Finally, from July to August the largest classes of oocyte size gradually recovered, the maximum average oocyte size being observed during the last month.Seventy one E. macha individuals of 1.6-23.5 mm L were found in the sediment samples.From these, spat smaller than 4 mm L were only present in November, July and August, suggesting that spawning may have taken place during spring and late fall, as was shown by the vitellogenic oocyte frequency distributions.

DISCUSSION
An abrupt change in growth pattern of ln(H) relative to ln(L), from negatively allometric to isometric, reveals that the valves of individuals smaller than 11 mm in L modify their shapes from more square to more elongated as they grow and that for larger individuals the shape does not vary with size.The elongated form of some infaunal species like razor clams enables them to burrow deep with low energy requirements to escape from predators (Trueman, 1967;Urban, 1994).For recently settled clam  spat, which are light and unable to burrow deep, stability against bottom currents may also be important for survival (Stanley, 1970).Since the form of E. macha spat has a lower surface-to-volume ratio than that of adults, it is probably more stable against the drag forces of bottom currents.As individuals grow, shell elongation might give them an adaptive advantage to become effective burrowers.Therefore, the change in the growth of H relative to L can be thought of as the result of a shift in the mode of life of individuals, from subsurface dwelling spat to active digging juveniles.The growth rates of TW and SPW relative to L found in E. macha in the present study are lower than those found for Ensis ensis and Ensis siliqua by Henderson and Richardson (1994).However, since the protocol used to weigh the last two species was different from that used for by us for E. macha, this comparison must be considered cautiously.The slightly negatively allometric growth of SPW relative to TW reveals that there is a slight loss in meat yield at the expense of shell weight as the clams grow.
Growth estimates obtained in the present study by observer 1 resulted in a von Bertalanffy's k value of 0.25 yr -1 , which is similar to that estimated by Carmona et al. (1999) by tracing modal components of size of an E. macha Chilean population.On the other hand, the parameter k obtained by Observer 2 was similar to that obtained by Urban (1996) by mark and recapture (k = 0.21 yr -1 ).L ∞ values obtained by both observers were lower than those estimated by Carmona et al. (1999) (216 mm) and Urban (1996) (189.9 mm).Differences in the estimates of k by both observers might be due to the interpretation of growth checks.Gaspar et al. (1994) analysed the growth checks of E. siliqua both by direct observation and using acetate peel replicas of shell sections, reaching the conclusion that growth check counting with the first method underestimated the growth rate (k) in that species.These authors noted that the clams that are disturbed form clefts on the surface of their valves that are frequently confused with seasonal growth checks under direct observation.On the other hand, there is another razor clam species (Ensis directus) in which seasonal growth checks are more conspicuous than disturb marks, allowing a reliable estimation of the growth parameters using direct observations (Swennen et al., 1985).The length of an individual at its 7 th growth check was 127.8 mm in the growth model of observer 1 and 120.9 mm in that of observer 2. Since the only marked individual recovered formed its 7 th growth check between 125.7 mm and 130.1 mm, its growth seems to be better explained by the first model.However, since there is a degree of variability in the length-at-growth check data (Table 2), no definitive conclusion can be derived on this aspect.Even though the validation of the periodicity of growth check deposition in E. macha was based on the observations conducted in one individual, the maximum number of growth checks found (observer 1: 14; observer 2: 16) is similar to the maximum age (13.5 y) estimated by Urban (1996) for the same species, giving additional support to the consideration of an annual deposition pattern for growth checks.
The present work shows that in northern Argentine Patagonia, the male to female ratio of E. macha is not significantly different from 1:1 for the three size classes analysed.This indicates that mortality throughout the size range of 90-165 mm is not sex dependent.In this population females do not have a resting period and they show two spawning peaks, one in spring (September to November) and the other in late autumn (May to June).Studies conducted by Avellanal et al. (2002) in the same species show that the Chilean populations within 37-42°S also remain in advanced stages of maturity (late active, ripe and partially spawned) most of the year.Furthermore, two peaks of maturation were observed by these authors in two out of three beds studied (Avellanal et al., 2002).Other razor clam species, such as E. siliqua from the coasts of Portugal (Gaspar and Monteiro, 1998), Solen strictus from Korea (Hong and Lee, 1990) and Siliqua patula from the Pacific coast of Canada (Bourne, 1979) spawn only once a year, generally from spring to summer.The samples of E. macha considered in this study came from populations located at the northernmost (warmest) limit of the species distribution.Specimens of this species and other clams with similar geographic distributions (e.g.Protothaca antiqua and Eurhomalea exalbida; Schuldt, 1975;Verdinelli and Schuldt, 1976;Borzone, 1989), are found in mature condition all year round.Also, they spawn more than once a year, either in the cold or the warm seasons.In contrast, other clam species whose southern (coldest) limit of distributions is the northern Patagonia (e.g.Tellina petitiana, Barón and Ciocco, 2001;Amiantis purpurata, Morsán, 2000) show clearly defined spawning peaks and mature individuals only during the late spring-summer period.
This study documents the general patterns of growth and reproduction for the northern Patagon-ian populations of E. macha.Although this information is relevant to regulate the imminent fishing activity, more detailed studies are still necessary to allow the design of a long-term management plan.However, preliminary exploitation strategies could be considered based on the present information.For example, brief reproductive closures could be implemented to avoid disturbance of the beds during spawning peaks.If this measure is to be taken, the annual schedule of male evacuation should be also analysed.Size limits could also be evaluated taking into account the biomass yield of a cohort under a reasonable range of natural mortality coefficients.

TABLE 3 .
-Sex ratio of Ensis macha divided into three length (L) classes